What is Distinctive about Cost-Benefit Analysis as a Policy Formulation Tool?
While the practical details can vary, the basic structure of any cost-benefit analysis takes the same form the world over. It involves summing the monetary value of net benefits (benefits minus costs) over the lifetime of a project or a policy. For a typical intervention, there will be costs now in return for benefits later. This leads to a crucial point: this stream of net benefits is discounted. What this means is that the value of net benefits in each period are not just added together but are treated differently depending on when they occur in time. Specifically, less and less weight is given to costs or benefits the further these impacts are in the future. These discounted net benefits are summed to estimate the net present value (NPV) of the project or policy. The decision rule in CBA is to recommend that the action goes ahead if the sum of (discounted) net benefits is greater than zero: that is, NPV>0. If we are choosing between mutually exclusive projects, then a CBA would recommend the project with the greatest net benefits.
The emphasis of conventional CBA is on securing overall net gains rather than their distribution. Placing this spotlight on total costs and total benefits does not necessarily reflect a judgement that distributional concerns 'do not matter'. Rather it assumes that the issue of who gains and who loses can be dealt with separately to the issue of making decisions, so as to generate as much overall 'goodness' of action as possible.2
Nevertheless, there are well-known procedures in CBA to deal with the benefits received and costs incurred by different societal groups. Indeed, in the UK, official guidance on CBA clearly states that:
Any distributional effects identified should be explicitly stated and quantified as far as possible. At a minimum, this requires appraisers to identify how the costs and benefits accrue to different groups in society .... Where it is considered necessary and practical, this might involve explicitly recognising distributional effects within a project's NPV. (HM Treasury 2003, p. 91)
Yet it is rare for such advice to be followed to the letter. Some deviation is understandable. What distributional weights should be applied is a matter of significant debate and uncertainty. Cost-benefit analysts can be forgiven perhaps for steering clear of these deep waters. However, it is harder to justify why relatively simple steps - such as identifying and cataloguing how costs and benefits are distributed - are seldom seen.
Another distributional dilemma surrounds discounting which, on the face of it, appears inconsistent with the spirit of 'intergenerational equity'. The reason is that the higher the discount rate used the more likely it is that (other things being equal) decisions are shifted towards actions which bring more immediate net benefits. Moreover, impacts occurring relatively far into the future receive almost no weight for any positive discount rate. Not surprisingly then, the choice of discount rate for policies with long-term consequences is especially controversial. For example, in the case of assessing the economic burden of climate change, this debate has highlighted fundamental differences between those economists who see a role for the practitioner to make, or reflect, explicit moral judgements about intergenerational equity and those who argue for a more objective approach based on information revealed in actual economic decisions about how much people care about the future (IPCC 1996; Stern 2007; Weitzman 2007). Resolving such debates is far from straightforward and faces profound questions on which, to quote Beckerman and Hepburn (2007, p. 198), 'reasonable minds may differ'. While it is important, therefore, that cost-benefit appraisal codifies and accommodates these differences, this may incur a penalty in the sense of less decisive recommendations.
Current interest in CBA stems from a variety of motivations. In part, however, the growing ability of practitioners to place robust money values on intangible impacts has surely played a crucial role. In environmental applications of CBA, for example, the estimation of these non-market values has given rise to a proliferation of methods.3 Some involve estimating original values by looking at actual behaviour. An example here would be the costs that visitors incur (in terms of out-of-pocket expenses and time costs of visits) to participate in nature-based recreation. Other methods elicit values by looking at intended behaviour. These stated preference methods involve people being asked more directly about the value they place on a policy change (see, for example, Alberini and Kahn 2006). Critical scrutiny of such techniques has also been prominent however. Grounds for criticism vary. For some, all non-market valuation appears to be controversial. Others make a distinction between certain classes of (non-market) goods which morally should be beyond valuation (for example, Kelman 2005), or technically defy robust valuation given the state-of-the-art in valuation practice (for example, Bateman et al. 2008; 2010). None of this criticism is necessarily a bad thing. A healthy dose of scepticism is important in the application, use and interpretation of any empirical methodology and non-market valuation is no exception.